Filtration of images and spatial trajectories using partial differential equations

The presented work is an attempt to combine two intensively developing fields: motion processing and image processing. In complicated vision systems, these areas often merge as successive stages of information processing. In this work, the analysis of connections is carried out at the level of data processing methodology: motivations related to similar assumptions and the use of opportunities resulting from the specificity of a given problem. An important task performed in the frame of image processing is filtration. The purpose of filtration is to extract elements from the image that are useful information in a given system.

Table of Contents

1. Introduction

2. Comparison of filtration methods for two-dimensional images

3. The proposal to include in the filtration process global knowledge about the geometric structure of the image

4. Models describing the spatial orientation of objects

5. Proposals of trajectory processing in S3 rotation spaces

6. The results of experimental research

7. Summary

Bibliography

List of illustrations

List of algorithms

List of markings